Read and analyze text files containing both numbers and letters on the same line

Question

I'm given a task to read data from a text file and save it to a Set. Text file represents an imaginary bill containing certain item description and it's price, quantity and total sum. I only need item's name and price. Text file looks like this:

Item_name   Item_price(float value with comma as format symbol)  Quantity(int)  Total(float)
Item_name   Item_price(float value with comma as format symbol)  Quantity(int)  Total(float)

(Text file contains multiple items). Also, items sometimes have numbers in their name, eg. LG 4k TV 1000U).

I tried to solve it like this:

private void readAndSave(Path file) {
    try (BufferedReader br = new BufferedReader(new InputStreamReader(
             new BufferedInputStream(new FileInputStream(file.toString()))))) {


        Set<Item> items = new TreeSet<>();
        String line;
        while ((line = br.readLine()) != null) {


            float price = 0, numb;
            boolean priceFound = false;
            String name = "";
            String[] lineElements;
            lineElements = line.split(" ");

            for(String temp: lineElements) {
                if((numb = getNumberRepresentation(temp)) != -1) {
                    if(!priceFound) {
                        price = numb;
                        priceFound = true;
                    }
                    break;
                }

                name += temp + " ";
            }
            items.add(new Item(name, price));

        }
    } catch (FileNotFoundException fe) {
        System.out.println("File not found!");
    } catch (IOException e) {
        System.out.println("Error while opening/writing files!");
    }

}

Class Item contains two variables(String, float) representing name and price of an item and extends Comparable.

And here is getNumberRepresentation method

private float getNumberRepresentation(String temp) {
    try {

        DecimalFormatSymbols symbols = new DecimalFormatSymbols();
        symbols.setDecimalSeparator(',');
        DecimalFormat format = new DecimalFormat("0.##");
        format.setDecimalFormatSymbols(symbols);
        return format.parse(temp).floatValue();

    } catch(Exception e) {
        return -1;
    }
}

I've tried to use the logic that, if a price is found, then the name must also be already found and all other Strings from the line can be skipped. Problem here is that sometimes I get a number from an item's name as price(1000U, from previous example). Is there a better and more efficient solution to this problem?

Edit: File sample

Escape from Paradise City 70,00 1135 79450,00 Sony ITC60, TV cabel 111,26 111 12349,86

One suggestion to your existing code... dont use 'try, catch' to parse you temp valiable... you should use TryParse() e.g. bool result = Int32.TryParse(value, out number); if (result) { Console.WriteLine("Converted '{0}' to {1}.", value, number); } else { // if (value == null) value = ""; Console.WriteLine("Attempted conversion of '{0}' failed.", value == null ? "" : value); } — Gwasshoppa, Dec 21 '16 at 00:52
Do you have control over what the text file looks like. i.e. Can you make the file TAB delimited, then you are splitting on a TAB not a " " (space. — Gwasshoppa, Dec 21 '16 at 00:57
Thanks for the suggestion. Unfortunately i don't have any control over the text file. — kolsdj, Dec 21 '16 at 01:07
Please add to the question a sample of the file lines if possible. — yaitloutou, Dec 21 '16 at 14:50

Jimi Kimble · Accepted Answer · 2016-12-21T17:53:37.927

0

You need to be using java.util.regex.Pattern and get everything before the match of a Regex for the cost and the match of the regex for the cost. I assume there won't be anything in the name that looks like ###,## where # are numerals. (Represented by \d in regex).

The tutorial can be found here.

It would look something like this:

Before reading the lines:

Pattern p = Pattern.compile("(.*?) (\\d*,\\d*)");

For each line:

Matcher m = p.matcher(line);
if (m.matches() && m.groupCount() == 2) {
    name = m.group(1);
    price = getNumberRepresentation(m.group(2));
} else {
    // line doesn't match the pattern, handle the exception!
}

edited Dec 21 '16 at 17:53

answered Dec 21 '16 at 01:48

Jimi Kimble

424
4
9

I've tried to use your approach(assuming that I understood it correctly), but for name it gets everything before the sum and for the price it gets the sum. Here is the code: ` lineElements = line.split(" "); Matcher m = p.matcher(line); if (m.matches() && m.groupCount() == 2) { name = m.group(0); price = getNumberRepresentation(m.group(1)); } else { System.out.println("Line error."); } ` – kolsdj Dec 21 '16 at 11:52
I realize now that my explanation is written poorly.. instead of saying before splitting the lines, I should have said before reading and for each line.. I'll fix the answer to reflect this. Also.. make sure your pattern is non-greedy. ".*?" instead of ".*". The first version of this answer was greedy which would give you the behavior you are talking about. – Jimi Kimble Dec 21 '16 at 17:52

Read and analyze text files containing both numbers and letters on the same line

1 Answers1